Filters with Small Non-Linearities
Abstract
The Kalman filter provides a finite dimensional solution when the signal and observation processes are linear and have Gaussian noise. In this paper the effect of a small nonlinearity in the signal is discussed by considering stochastic flows for the signal and a Girsanov transformation for the observation. The result can be expressed in terms of Gaussian densities.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 01, 1988
- Accession Number
- ADA222215
Entities
People
- Robert J. Elliott
Organizations
- University of Alberta